Re: Further Thoughts on the Origin of Chelonia

["Tim Williams" <twilliams_alpha@hotmail.com>]

Thomas R. Holtz, Jr. wrote:

> Incidentally, I think that (much as I would like turtles to be surviving
> pareiasaurs) they are going to prove to be archosauromorphs. The molecules
> really do seem to point that way, particularly as the taxon sampling has
> increased (so it isn't just one croc, one bird, one squamate, and one
> turtle).

This is not in contradiction to Tom - after all, I also happen to believe 
that our shelled friends (including my pet turtle) are secondarily anapsid, 
and probably some sort of diapsid.  However, improved taxon sampling does 
not always improve the resolution of a given phylogeny.  Molecules have 
quirks and idiosyncracies, just as morphological characters do - DNA just 
happens to have different sorts of quirks and idiosyncracies.  Expanding the 
number of taxa, and/or expanding the number of genes or sequences, does not 
automatically 'smooth over' the bumps in a molecular-based phylogeny.

The usefulness of a gene in capturing deep divergences (as we have with 
turtles vs other amniotes) relies on the gene having evolved slowly enough 
so as not to become saturated with substitutions, but not too slowly that 
too few substitutions accumulate over time (such as if the encoded protein 
is functionally constrained from evolving too much).  Expanding the taxon 
sample does not in of itself overcome this problem.  Also, even within a 
given clade, the same genes from different species can have their own 
intrinsic biases (e.g., codon usage, GC content), which can further mess 
things up, since these biases can be mistaken for synapomorphies.  Thus, 
unrelated critters can be pulled together on account of these biases.

I'm not saying that the phylogeny or phylogenies that Tom is referring to 
are incorrect.  But, theoretically (and this is a very hard thing to do), if 
one could remove the "noise" that is disrupting the evolutionary signal, 
then a smaller number of taxa would suffice to produce an accurate 
phylogeny.



Tim



Tim

_________________________________________________________________
Send instant messages to anyone on your contact list with  MSN Messenger 
6.0.  Try it now FREE!  http://msnmessenger-download.com